Submitted 11 June 2016Accepted 27 September 2016Published 26 October 2016

Corresponding authorJonathan Lombard,lombard.jonathan@yahoo.com,j.lombard@exeter.ac.uk

Academic editorGilles van Wezel

Additional Information andDeclarations can be found onpage 10

DOI 10.7717/peerj.2626

Copyright2016 Lombard

Distributed underCreative Commons CC-BY 4.0

OPEN ACCESS

Early evolution of polyisoprenolbiosynthesis and the origin of cell wallsJonathan LombardBiosciences, University of Exeter, Exeter, United KingdomNational Evolutionary Synthesis Center, Durham, NC, United States of America

ABSTRACTAfter being a matter of hot debate for years, the presence of lipid membranes in thelast common ancestor of extant organisms (i.e., the cenancestor) now begins to begenerally accepted. By contrast, cenancestral cell walls have attracted less attention,probably owing to the large diversity of cell walls that exist in the three domains of life.Many prokaryotic cell walls, however, are synthesized using glycosylation pathwayswithsimilar polyisoprenol lipid carriers and topology (i.e., orientation across the cell mem-branes).Here, we provide the first systematic phylogenomic report on the polyisoprenolbiosynthesis pathways in the three domains of life. This study shows that, whereas thelast steps of the polyisoprenol biosynthesis are unique to the respective domain of life ofwhich they are characteristic, the enzymes required for basic unsaturated polyisoprenolsynthesis can be traced back to the respective last common ancestor of each of the threedomains of life. As a result, regardless of the topology of the tree of life that may beconsidered, the most parsimonious hypothesis is that these enzymes were inherited inmodern lineages from the cenancestor. This observation supports the presence of anenzymatic mechanism to synthesize unsaturated polyisoprenols in the cenancestor and,since these molecules are notorious lipid carriers in glycosylation pathways involved inthe synthesis of a wide diversity of prokaryotic cell walls, it provides the first indirectevidence of the existence of a hypothetical unknown cell wall synthesis mechanism inthe cenancestor.

Subjects Biochemistry, Evolutionary Studies, Genomics, MicrobiologyKeywords Polyisoprenol, Lipid carrier, Cenancestor, LUCA, Cell walls, Glycosylation, Archaea,Bacteria, Eukaryotes

INTRODUCTIONCells from the three domains of life (archaea, bacteria and eukaryotes) are bound by cellmembranes of which the main lipid components are the phospholipids. At first, the strongchemical and biosynthetic dissimilarities that exist between the archaeal phospholipids andthose from bacteria and eukaryotes triggered a hot debate about the existence of lipid mem-branes in the last common ancestor of extant organisms, namely the cenancestor or LUCA(Koga et al., 1998; Wächtershäuser, 2003; Martin & Russell, 2003; Peretó, López-García& Moreira, 2004). Later, new phylogenomic analyses inferred the presence of metabolicpathways responsible for the synthesis ofmost phospholipid components in the cenancestor(Peretó, López-García & Moreira, 2004; Lombard & Moreira, 2011; Lombard, López-García& Moreira, 2012a; Lombard, López-García & Moreira, 2012c). Therefore, the debates have

How to cite this article Lombard (2016), Early evolution of polyisoprenol biosynthesis and the origin of cell walls. PeerJ 4:e2626; DOI10.7717/peerj.2626

progressively shifted from arguing the existence of membranes to discuss their compositionand properties (Lane & Martin, 2012; Lombard, López-García & Moreira, 2012b; Sojo,Pomiankowski & Lane, 2014). Contrary to cell membranes, the early evolution of cell wallshas been scarcely studied. Cell walls indeed provide essential structural support and externalinteractions inmodern organisms (Albers & Meyer, 2011), but their extremediversitymakesthe study of their early evolution very challenging.

Despite the stunning diversity that exists among prokaryotic cell envelopes, thesynthesis of many of their main components (e.g., N -or O-glycosylated S-layer proteins,peptidoglycan, O-antigen LPS, teichoic acids, exopolysaccharides) relies on comparableglycosylation pathways (Hartley & Imperiali, 2012; Lombard, 2016). These pathways are alllocated in the cell membranes, are mediated by similar lipid carriers and have the sameorientation across the membrane. Their respective mechanisms can be described in threemajor steps: (1) synthesis of an oligomer on a lipid carrier in the cytoplasmic side of thecell membranes; (2) ‘‘flipping’’ of the oligomer-linked lipid carrier to the opposite side ofthe membrane; and (3) oligomer transfer from the lipid carrier to the acceptor molecule(e.g., protein or external polymer). The eukaryotic proteinN -glycosylation alsomeets thesecriteria, although it takes place in the ER membranes instead of the plasma membranes.The proteins involved in these glycosylation pathways belong to a small number ofprotein superfamilies, some of which could be traced back to the cenancestor (Lombard,2016). The large diversity of the glycosylation pathways makes it difficult to untangle thespecific evolutionary relationships that exist among them but, despite their differences, theprokaryotic cell wall synthesis mechanisms have at least one element in common: they alluse polyisoprenol phosphate lipid carriers (Hartley & Imperiali, 2012).

Polyisoprenol phosphates are long unsaturated chains with a varying number of isopreneunits and an α-terminal phosphate group (Fig. 1). The eukaryotic polyisoprenol phosphateis called dolichol-phosphate (Dol-P) and it is characterized by the saturation of its α-unit. Archaeal lipid carriers are α- and ω-polysaturated and also called Dol-P. Bacterialpolyisoprenols are fully unsaturated and their names vary according to their length (Meyer& Albers, 2013), so they will be referred to as bactoprenol phosphate (Bac-P) for simplicity.

Similar to other isoprenoids, the polyisoprenols are initially synthesized through thesequential condensation of isopentenyl pyrophosphate (IPP) and allylic prenyl diphos-phates, the first of which is dimethylallyl pyrophosphate (DMAPP, Fig. 1). The first 3–4condensations in polyisoprenol biosynthesis require regular E-prenyltransferases (IPPS)but, contrary to other isoprenoids, later elongations involveZ -type links (Ogura & Koyama,1998; Schenk, Fernandez & Waechter, 2001;Guan et al., 2011). The rest of the polyisoprenolphosphate synthesis pathways differs in each domain of life (Fig. 1). In bacteria, polyprenoldiphosphate must be dephosphorylated into Bac-P. Most (75%) of this activity is catalyzedby an undecaprenyl pyrophosphate phosphatase (UppP/BacA) (El Ghachi et al., 2004;Chang et al., 2014; Manat et al., 2015), and complemented by the members of a promis-cuous family of PAP-motif phosphatases (Stukey & Carman, 1997; El Ghachi et al., 2005;Touzé, Blanot & Mengin-Lecreulx, 2008). The involvement in this dephosphorylation step ofother components, such as a supplementary cytoplasmic phosphatase or a polyprenoldiphosphate translocation mechanism, has been suggested (see below), but their identity

Lombard (2016), PeerJ, DOI 10.7717/peerj.2626 2/15

Figure 1 (A) Typical polyisoprenol lipid carriers in the three domains of life. (B) Polyisoprenol-phosphate biosynthesis pathways in the three domains of life. Steps outside boxes are common to allpathways, whereas steps in boxes are characteristic of only one domain of life. Unknown steps are labeledwith a question mark. GPPS, geranyl diphosphate synthase; FPPS, farnesyl diphosphate synthase; GGPPS,geranylgeranyl diphosphate synthase; GGR, geranylgeranylreductase; IPT, isopentenyl transferase; UppP,undecaprenyl pyrophosphate phosphatase; PR, polyprenol reductase; DK, dolichol kinase.

Lombard (2016), PeerJ, DOI 10.7717/peerj.2626 3/15

remains unknown. In archaea, the ω-terminal and internal saturated units are reducedby the geranylgeranyl reductases (GGR, Fig. 1) either before the elongation of FPP/GGPPinto Dol-P or later on (Sato et al., 2008; Guan et al., 2011; Naparstek, Guan & Eichler,2012). The rest of the archaeal Dol-P biosynthesis pathway remains unknown. Finally,in eukaryotes, polyprenol diphosphate is dephosphorylated, α-reduced (Cantagrel et al.,2010) and rephosphorylated (Cantagrel & Lefeber, 2011, Fig. 1). Here, a phylogenomicanalysis of the proteins involved in the polyisoprenol phosphate biosynthesis pathways wasperformed in the three domains of life. This analysis provides promising insights into theunexplored question of the early evolution of cell walls.

MATERIAL AND METHODSThe original sequence seeds were selected from KEGG (http://www.genome.jp/kegg/) andverified in the literature. BLASTp (Altschul et al., 1990) were carried out against a selectionof representative genomes (Fig. S1) or the non-redundant (nr) dataset on GenBank(http://www.ncbi.nlm.nih.gov/Genbank), depending on results. When homologues weredifficult to detect, more distant homologues were searched for using PSI-BLAST (Altschulet al., 1997) or HMMprofiles and the hmmsearch tool implemented in the HMMER v3.1b2webserver (http://hmmer.org/download.html and Finn et al., 2015). The alignments werecarried out with MUSCLE v3.8.31 (Edgar, 2004) and trimmed with the program NET ofthe MUST package (Philippe, 1993). The preliminary phylogenies were reconstructed usingFastTree v.2.1.7 (Price, Dehal & Arkin, 2010). Representative sequences were selected fromthese phylogenies to carry out more accurate analyses, and realigned usingMUSCLE on theGUIDANCE server (http://guidance.tau.ac.il/ver2/, Penn et al., 2010). The final alignmentsare provided in Fig. S2 and were trimmed based on the statistical scores provided byGUIDANCE. The final phylogenies were constructed using MrBayes 3.2.6 (Ronquist et al.,2012) (LG substitution model (Le & Gascuel, 2008); 4 0 categories; 4 chains of 1,000,000generations; tree sampling every 100 generations; 25% generations discarded as ‘‘burn in’’)and RaxML-HPC2 8.1.24 (Stamatakis, 2014) (LG + 0 model (Le, Dang & Gascuel, 2012);4 rate categories; 100 bootstrap replicates) implemented in the CIPRES Science Gateway(Miller, Pfeiffer & Schwartz, 2010). Bayesian and maximum likelihood phylogenies gavecomparable results.

RESULTS AND DISCUSSIONDespite being widespread in archaea, bacteria and eukaryotes, IPP and DMAPP aresynthesized by respective pathways in each domain of life (Lombard & Moreira, 2011;Dellas et al., 2013; Vranová, Coman & Gruissem, 2013). A previous study on the evolutionof these pathways concluded that the cenancestor had a primitive mevalonate pathway thatcould have synthesized IPP and DMAPP in this organism, and that each domain of lifeadopted its specific biosynthesis pathway later on (Lombard & Moreira, 2011). The E-typeprenyl condensations are carried out by E-prenyltransferases (IPPS). The evolution of theseproteins has been described in prokaryotes (Lombard, López-García & Moreira, 2012a) andthe analysis is now extended to eukaryotes (see Fig. S3 for details). In summary, these

Lombard (2016), PeerJ, DOI 10.7717/peerj.2626 4/15

analyses are not conclusive on the specific relationships that may exist among IPPS fromthe three domains of life, but they support the presence of IPPS homologues in the respectivecommon ancestor of each domain of life. As a result, their presence can also be inferredin the cenancestor independently of the topology of the tree of life that may be favored.

The most characteristic enzyme of the polyisoprenol phosphate biosynthesis pathwayis the Z -isopentenyltransferase (Z -IPTase, Fig. 1) (Shimizu, Koyama & Ogura, 1998). Thephylogeny of the Z -IPTases (Fig. 2) shows archaeal and bacterial monophyletic clades(Bayesian posterior probability (BPP) = 0.77 and BPP = 1, respectively) and severalgroups of eukaryotic sequences. The internal relationships within the archaeal clade arepoorly resolved, which is not unexpected for a single-gene tree. The euryarchaea areparaphyleticand contain a group made up of distant paralogues and bacterial sequencesprobably acquired through horizontal gene transfer (HGT). Nevertheless, Z -IPTasehomologues are widespread among archaea and the monophyly of the archaeal andproteoarchaeal (TACK superphylum, Petitjean et al., 2015) sequences support the presenceof this gene in the Last Archaeal Common Ancestor. The deep phylogenetic relationshipswithin the bacterial clade are also unresolved. Yet, the wide diversity of bacteria in whichZ -IPTase homologues may be detected, together with the fact that most bacterial sequencesgroup according to their taxonomic classification supports the ancestral presence of thisgene in bacteria. Some polyphyletic sequences from plastid-bearing eukaryotes branchwithin a group of cyanobacteria and proteobacteria (BPP= 0.88) and likely have plastidialancestors. The rest of the eukaryotic sequences are paraphyletic. The largest eukaryoticclade (BPP= 1) is made up of diverse sequences that cluster according to major eukaryotictaxa (e.g., opisthokonts, streptophytes, alveolates). Some internal branching is suggestiveof inter-eukaryotic HGT, such as in organisms known to have had secondary plastidendosymbioses, which branch among archaeplastida. A smaller cluster of divergenteuglenozoan and stramenopile sequences (BPP = 1) forms a sister group to the largeeukaryotic/archaeal clade. Some of euglenozoa are known to synthesize shorter Dol-Pmolecules than other eukaryotes, so it is possible that the phylogenetic position of thesesequences reflects a reconstruction artifact due to their high divergence (Schenk, Fernandez& Waechter, 2001). Ultimately, if we accept that the few paraphylies and polyphylies inthis tree result from reconstruction artifacts due to the extreme sequence divergence ofZ -IPTases across the three domains of life, this (unrooted) phylogeny supports the presenceof Z -IPTases in the respective last common ancestors of each domain of life and, regardlessof the topology of tree of life that may be invoked, in the cenancestor.

The following steps in the polyisoprenol phosphate biosynthesis pathways are specific toeach domain of life (Fig. 1). In bacteria, the polyprenol diphosphate is dephosphorylatedinto Bac-P by UppP (BacA) or a family of periplasmic PAP-motif phosphatases. UppPand the PAP-motif phosphatases are not evolutionary related to each other and, untilrecently, it was thought that they were respectively responsible for the cytoplasmic de novosynthesis or the periplasmic recycling of polyprenol diphosphate (Bickford & Nick, 2013;Manat et al., 2014). The recent discovery that they all function in the periplasmic side of themembrane implies that the de novo polyprenol diphosphate dephosphorylation requireseither the activity of an unknown cytoplasmic phosphatase, or the translocation to the

Lombard (2016), PeerJ, DOI 10.7717/peerj.2626 5/15

Figure 2 Bayesian phylogeny of the Z -IPTase homologues. The tree is unrooted and was reconstructedusing 197 representative sequences and 202 conserved sites. Multifurcations correspond to branches withBayesian posterior probabilities<0.5, whereas numbers at nodes indicate Bayesian posterior probabilitieshigher than 0.5. The bootstrap values from the maximum likelihood analyses have been reported on basaland major nodes. Colors of leaves represent the affiliation of sequences to their respective domain of life:archaea (blue), bacteria (orange) and eukaryotes (purple).

Lombard (2016), PeerJ, DOI 10.7717/peerj.2626 6/15

periplasmic side of the membrane through an unknown mechanism (Chang et al., 2014;Manat et al., 2015; Teo & Roper, 2015). Although some candidates have been suggested tomediate the polyprenol (di)phosphate translocation (Sanyal & Menon, 2010; Chang et al.,2014;Manat et al., 2015), these hypotheses require biochemical verification. Thus, only thetwo known phosphatase families will be studied here.

UppP homologues were detected in archaea and a wide diversity of bacteria but not ineukaryotes, even when HMM searches for distantly related sequences were carried out (see‘Methods’). Since relatively little is known about the last steps in archaeal polyisoprenolbiosynthesis, the possible participation of these archaeal homologues in this pathway seemsworth testing experimentally. The phylogeny of the UppP homologues is discussed in moredetail in Fig. S4 but, in summary, the respective monophyly of archaeal and bacterial UppPhomologues (BPP= 0.96) suggests that this gene could have been present in the respectiveancestors of both archaea and bacteria and, therefore, in the cenancestor. Importantly, thissuggests that the cenancestor could have had a minimum set of proteins representative ofall the enzymes currently known to be required to synthesize not only polyisoprenol chainsbut also functional lipid carriers (Bac-P). Regarding the PAP-motif phosphatases, severalrepresentatives of the family have been biochemically characterized (Fernández & Rush,2001; El Ghachi et al., 2005; Manabe et al., 2009) and their homologues are widespread inthe three domains of life, but their phylogeny is poorly resolved and dominated by thepresence of paralogues and xenologues (see Fig. S5 for details). As a result, it is not possibleat this stage to discriminate between a cenancestral or more recent origin of the PAP-motifphosphatases.

In archaea, the ω-terminal and internal units are reduced by the GGRs (Fig. 1). Arecent phylogenomic analysis of the GGRs suggested (1) the ancestral presence of thisgene in archaea, followed by a complicated history of duplications and HGTs; (2) anunclear origin for these genes in bacteria; and (3) a likely plastidial origin of these genes ineukaryotes (Lombard, López-García & Moreira, 2012a); thus, the presence of GGRs in thecenancestor is not supported. The rest of the archaeal Dol-P biosynthesis pathway remainsuncharacterized, so particular attention was drawn to the possible archaeal homologues ofthe enzymes in the eukaryotic Dol-P synthesis pathway.

The specific steps in the eukaryotic pathway consist in the dephosphorylation, α-reduction and rephosphorylation of the polyprenol diphosphate (Fig. 1). The eukaryoticpolyprenol diphosphate phosphatase is unknown (Cantagrel & Lefeber, 2011; Bickford &Nick, 2013). Sequences similar to the polyisoprenol α-unit reductases (PR, Fig. 1) arewidespread in eukaryotes but they are absent in archaea and rare in bacteria (these arelikely xenologues). This implies that archaea use an alternative way to saturate the α-unit oftheir Dol-P. The PR phylogeny (see Fig. S6 for details) suggests that this protein is an earlyeukaryotic innovation that was developed before the Last Eukaryotic Common Ancestorand duplicated later in several eukaryotic lineages. Finally, homologues of the dolicholkinase (DK) were easily detected among a large diversity of eukaryotes, but only in a fewprokaryotes. PSI-BLAST searches (Altschul et al., 1997) revealed a distant evolutionaryrelationship between the DKs and the CDP-diacylglycerol synthases (CdsA) involved inphospholipid synthesis (Sparrow & Raetz, 1985). The evolution of both functions was

Lombard (2016), PeerJ, DOI 10.7717/peerj.2626 7/15

Figure 3 Summary of the phylogenomic results of the enzymes involved in the polyisoprenol biosyn-thesis pathways.Only known genes are presented, so the eukaryotic and archaeal pathways are not com-plete. MVA, mevalonate; MEP, Methylerythritol phosphate; LACA, Last Archaeal Common Ancestor;LBCA, Last Bacterial Common Ancestor; LECA, Last Eukaryotic Common Ancestor. Refer to Fig. 1 forother abbreviations.

studied (Fig. S7) but the quality of the sequence alignment was very poor, the resultingphylogenies are unreliable and so are the conclusions that can be taken from them.This analysis, however, questions the existence of DK functional orthologues in mostprokaryotes, especially in Euryarchaea. The lack of homologues of the last steps ofeukaryotic Dol-P synthesis in most archaeal genomes raises the question of how theseorganisms finish the synthesis of their polyisoprenol phosphate lipid carriers.

CONCLUSIONSThe initial metabolic steps of the polyisoprenol biosynthesis pathways (from IPP/DMAPPsynthesis to polyprenol diphosphate) are widespread in the three domains of life, and theirpresence can be inferred in at least the respective ancestor of each prokaryotic domain oflife (Fig. 3). The later steps differ in each domain. Bacterial UppP (BacA) may be ancestralto both bacteria and archaea, but the evolution of the PAP-motif phosphatases is less clear.PR, and probably DK as well, are eukaryotic innovations. Archaea use their ancestral GGR,

Lombard (2016), PeerJ, DOI 10.7717/peerj.2626 8/15

they do not have PR homologues andmost of them lack DK homologues. Although neitherthe archaeal nor the eukaryotic pathways are fully described yet, the available informationsuggests that these domains have different means of synthesizing their polyisoprenolphosphates. This is congruent with the fact that, despite having the same name, eukaryoticDol-P and archaeal Dol-P are chemically different from each other (Fig. 1). In order toavoid any confusion, we propose the archaeal polysaturated polyisoprenols (Guan et al.,2010; Guan et al., 2011; Chang et al., 2015) to be called ‘‘archaeoprenols’’ from now on.

Polyprenol diphosphate biosynthesis is ancestral to each domain of life and, therefore, themost parsimonious hypothesis to explain its distribution is that it was inherited from thecenancestor. Although the mechanism that provides this molecule to the periplasmicphosphatases remains unknown (Chang et al., 2014;Manat et al., 2015), the possibility thatUppP homologuesmay also be ancestral to both prokaryotic domains similarly suggests thatthe cenancestor could even have been able to synthesize some sort of Bac-P. Polyisoprenolsmay carry out pleiotropic functions (Cantagrel & Lefeber, 2011; Hartley & Imperiali, 2012),but theirmost notorious role is as lipid carriers in a number of glycosylation pathways (Lom-bard, 2016). In prokaryotes, these glycosylations are systematically related to the synthesisof cell wall components, so it is reasonable to postulate that hypothetical cenancestralpolyisoprenols may have played the role of lipid carriers in early glycosylation mechanisms,possibly related to cell wall synthesis. Yet, this hypothesis will need to be updated whenmore information about the polyisoprenol biosynthesis pathways will be available.

Although the presence of a polyisoprenol phosphate biosynthesis pathway in thecenancestor is a promising piece of evidence to support the existence of cell walls in thisorganism, the large diversity of microbial cell walls remains an obstacle to risk a hypothesison the specific nature of this cenancestral structure. Interestingly, the presence of at leasttwo promising glycosyltransferase superfamilies (HPT and MurG) has been inferred in thecenancestor (Lombard, 2016). The proteins in these families are responsible for the transferof soluble monosaccharides to polyisoprenol phosphate lipid carriers in bacterial peptido-glycan synthesis (Mohammadi et al., 2007) and archaeal S-layer protein N -glycosylation(Meyer & Albers, 2013). If we sum up the results presented here and those about glycosyl-transferases (Lombard, 2016), it is tempting to say that the cenancestor could have beenable to use its HPT andMurG homologues to synthesize an oligosaccharide on a Bac-P lipidcarrier. The question that remains open is the final acceptor of that hypothetical cenancestraloligosaccharide, as that is the element that would define the nature of the putativecenancestral cell wall. One way to tackle this question would be studying the transferasesthat convey the glycans from the lipid carrier to various donors. These transferases werephylogenomically analyzed in (Lombard, 2016), but none of them revealed a widespreadfunction that would provide an obvious candidate to make up a particular type of cellwall in the cenancestor. For instance, S-layers have been put forward as possible firstcell wall components (Albers & Meyer, 2011; Fagan & Fairweather, 2014). The protein N -glycosylation oligosaccharyltransferases and O-mannosylation transferases responsible forS-layer protein glycosylation are indeed probably ancestral to eukaryotes and archaea, buttheir ancestral presence in bacteria is uncertain (Lombard, 2016). Therefore, the presenceof glycosylated S-layers proteins in the cenancestor remains inconclusive.

Lombard (2016), PeerJ, DOI 10.7717/peerj.2626 9/15

It has been argued that the strong conservation of polyisoprenol phosphates as lipidcarriersmay be functional: theymay play important regulatory roles on glycanmechanisms,their biochemistry may be particularly adapted to glycan translocation across phospholipidbilayers or they can be beneficial to the formation of large protein complexes with whichthey frequently interact (Hartley & Imperiali, 2012). Contrary to these possible strongpressures to maintain polyisoprenol phosphates as lipid carriers, the difficulty to determinethe final acceptor of the hypothetical cenancestral oligosaccharide may be explained by thelikely strong selective pressures that acted on the exposed cell wall components to diversify.For instance, one specific type of cell wall may have actually existed in the cenancestor, butthe strong selective pressures exerted on the cell wall glycocalyx may have resulted in theindependent replacement of the original cell wall in many lineages. Another possibility isthat the cenancestral cell wall glycosylation was promiscuous and that specific cell wall typesonly evolved later in modern lineages, following the same selective pressures. We are at atime of dynamic characterization of prokaryotic envelopes and comparative studies, so newevidence may soon shed light on more mechanistic similarities among cell wall synthesesin the three domains of life and complement our knowledge of the cenancestral envelopes.

ACKNOWLEDGEMENTSI thank U Gophna for the stimulating conversations that crystallized in the developmentof this project and D Moreira and TA Richards for their advice at later steps. I also thankD Moreira, TA Richards, C Petitjean, F Savory, R Bamford, V Attah, five reviewers fromprevious submissions and two reviewers from this submission for their comments on themanuscript.

ADDITIONAL INFORMATION AND DECLARATIONS

FundingThis work was supported by the National Science Foundation through the NationalEvolutionary Synthesis Center (grant number NSF #EF-0905606). The funders had no rolein study design, data collection and analysis, decision to publish, or preparation of themanuscript.

Grant DisclosuresThe following grant information was disclosed by the author:National Science Foundation: NSF #EF-0905606.

Competing InterestsThe authors declare there are no competing interests.

Author Contributions• Jonathan Lombard conceived and designed the experiments, performed the experiments,analyzed the data, contributed reagents/materials/analysis tools, wrote the paper,prepared figures and/or tables, reviewed drafts of the paper.

Lombard (2016), PeerJ, DOI 10.7717/peerj.2626 10/15

Data AvailabilityThe following information was supplied regarding data availability:

The raw data has been supplied as Supplementary File.

Supplemental InformationSupplemental information for this article can be found online at http://dx.doi.org/10.7717/peerj.2626#supplemental-information.

REFERENCESAlbers S-V, Meyer BH. 2011. The archaeal cell envelope. Nature Reviews Microbiology

9:414–426 DOI 10.1038/nrmicro2576.Altschul SF, GishW,MillerW,Myers EW, Lipman DJ. 1990. Basic local alignment

search tool. Journal of Molecular Biology 215:403–410DOI 10.1016/S0022-2836(05)80360-2.

Altschul SF, Madden TL, Schäffer AA, Zhang J, Zhang Z, MillerW, Lipman DJ. 1997.Gapped BLAST and PSI-BLAST: a new generation of protein database searchprograms. Nucleic Acids Research 25:3389–3402 DOI 10.1093/nar/25.17.3389.

Bickford JS, Nick HS. 2013. Conservation of the PTEN catalytic motif in the bac-terial undecaprenyl pyrophosphate phosphatase, BacA/UppP.Microbiology159:2444–2455 DOI 10.1099/mic.0.070474-0.

Cantagrel V, Lefeber DJ. 2011. From glycosylation disorders to dolichol biosynthesisdefects: a new class of metabolic diseases. Journal of Inherited Metabolic Disease34:859–867 DOI 10.1007/s10545-011-9301-0.

Cantagrel V, Lefeber DJ, Ng BG, Guan Z, Silhavy JL, Bielas SL, Lehle L, Hombauer H,Adamowicz M, Swiezewska E, De Brouwer AP, Blümel P, Sykut-CegielskaJ, Houliston S, Swistun D, Ali BR, DobynsWB, Babovic-Vuksanovic D,Van Bokhoven H. 2010. SRD5A3 is required for converting polyprenol todolichol and is mutated in a congenital glycosylation disorder. Cell 142:203–217DOI 10.1016/j.cell.2010.06.001.

Chang H-Y, Chou C-C, HsuM-F,Wang AHJ. 2014. Proposed carrier lipid-binding siteof undecaprenyl pyrophosphate phosphatase from Escherichia coli. The Journal ofBiological Chemistry 289:18719–18735 DOI 10.1074/jbc.M114.575076.

ChangMM, Imperiali B, Eichler J, Guan Z. 2015. N-Linked glycans are assembledon highly reduced dolichol phosphate carriers in the hyperthermophilic archaeaPyrococcus furiosus. PLoS ONE 10:e0130482 DOI 10.1371/journal.pone.0130482.

Dellas N, Thomas ST, Manning G, Noel JP. 2013. Discovery of a metabolic alternative tothe classical mevalonate pathway. ELife 2:e00672 DOI 10.7554/eLife.00672.

Edgar RC. 2004.MUSCLE: multiple sequence alignment with high accuracy and highthroughput. Nucleic Acids Research 32:1792–1797 DOI 10.1093/nar/gkh340.

Fagan RP, Fairweather NF. 2014. Biogenesis and functions of bacterial S-layers. NatureReviews Microbiology 12:211–222 DOI 10.1038/nrmicro3213.

Fernández F, Rush JS, Toke DA, Han G-S, Quinn JE, Carman GM, Choi J-Y, VoelkerDR, Aebi M,Waechter CJ. 2001. The CWH8 gene encodes a dolichyl pyrophosphate

Lombard (2016), PeerJ, DOI 10.7717/peerj.2626 11/15

phosphatase with a luminally oriented active site in the endoplasmic reticulumof Saccharomyces cerevisiae. Journal of Biological Chemistry 276:41455–41464DOI 10.1074/jbc.M105544200.

Finn RD, Clements J, ArndtW,Miller BL,Wheeler TJ, Schreiber F, Bateman A, EddySR. 2015.HMMER web server: 2015 update. Nucleic Acids Research 43:W30–W38DOI 10.1093/nar/gkv397.

El Ghachi M, Bouhss A, Blanot D, Mengin-Lecreulx D. 2004. The bacA gene of Es-cherichia coli encodes an undecaprenyl pyrophosphate phosphatase activity. Journalof Biological Chemistry 279:30106–30113 DOI 10.1074/jbc.M401701200.

El Ghachi M, Derbise A, Bouhss A, Mengin-Lecreulx D. 2005. Identification of multiplegenes encoding membrane proteins with undecaprenyl pyrophosphate phosphatase(UppP) activity in Escherichia coli. Journal of Biological Chemistry 280:18689–18695DOI 10.1074/jbc.M412277200.

Guan Z, Meyer BH, Albers S-V, Eichler J. 2011. The thermoacidophilic archaeon Sul-folobus acidocaldarius contains an unsually short, highly reduced dolichyl phosphate.Biochimica et Biophysica Acta 1811:607–616 DOI 10.1016/j.bbalip.2011.06.022.

Guan Z, Naparstek S, Kaminski L, Konrad Z, Eichler J. 2010. Distinct glycan-chargedphosphodolichol carriers are required for the assembly of the pentasaccharideN-linked to the Haloferax volcanii S-layer glycoprotein.Molecular Microbiology78:1294–1303 DOI 10.1111/j.1365-2958.2010.07405.x.

Hartley M, Imperiali B. 2012. At the membrane frontier: a prospectus on the remarkableevolutionary conservation of polyprenols and polyprenyl-phosphates. Archives ofBiochemistry and Biophysics 517:83–97 DOI 10.1016/j.abb.2011.10.018.

Koga Y, Kyuragi T, Nishihara M, Sone N. 1998. Did archaeal and bacterial cells ariseindependently from noncellular precursors? A hypothesis stating that the advent ofmembrane phospholipid with enantiomeric glycerophosphate backbones causedthe separation of the two lines of descent. Journal of Molecular Evolution 46:54–63DOI 10.1007/PL00006283.

Lane N, MartinWF. 2012. The origin of membrane bioenergetics. Cell 151:1406–1416DOI 10.1016/j.cell.2012.11.050.

Le SQ, Dang CC, Gascuel O. 2012.Modeling protein evolution with several aminoacid replacement matrices depending on site rates.Molecular Biology and Evolution29:2921–2936 DOI 10.1093/molbev/mss112.

Le SQ, Gascuel O. 2008. An improved general amino acid replacement matrix.MolecularBiology and Evolution 25:1307–1320 DOI 10.1093/molbev/msn067.

Lombard J. 2016. The multiple evolutionary origins of the eukaryotic N-glycosylationpathway. Biology Direct 11:36 DOI 10.1186/s13062-016-0137-2.

Lombard J, López-García P, Moreira D. 2012a. Phylogenomic investigation of phospho-lipid synthesis in archaea. Archaea 2012:630910 DOI 10.1155/2012/630910.

Lombard J, López-García P, Moreira D. 2012b. The early evolution of lipid mem-branes and the three domains of life. Nature Reviews Microbiology 10:507–515DOI 10.1038/nrmicro2815.

Lombard (2016), PeerJ, DOI 10.7717/peerj.2626 12/15

Lombard J, López-García P, Moreira D. 2012c. An ACP-independent fatty acid synthesispathway in archaea: implications for the origin of phospholipids.Molecular Biologyand Evolution 29:3261–3265 DOI 10.1093/molbev/mss160.

Lombard J, Moreira D. 2011. Origins and early evolution of the mevalonate pathway ofisoprenoid biosynthesis in the three domains of life.Molecular Biology and Evolution28:87–99 DOI 10.1093/molbev/msq177.

Manabe F, Itoh YH, Shoun H,Wakagi T. 2009.Membrane-bound acid pyrophosphatasefrom Sulfolobus tokodaii, a thermoacidophilic archaeon: heterologous expres-sion of the gene and characterization of the product. Extremophiles 13:859–865DOI 10.1007/s00792-009-0273-z.

Manat G, El Ghachi M, Auger R, Baouche K, Olatunji S, Kerff F, Touzé T, Mengin-Lecreulx D, Bouhss A. 2015.Membrane topology and biochemical characterizationof the Escherichia coli BacA undecaprenyl-pyrophosphate phosphatase. PLoS ONE10:e0142870 DOI 10.1371/journal.pone.0142870.

Manat G, Roure S, Auger R, Bouhss A, Barreteau H, Mengin-Lecreulx D, Touzé T.2014. Deciphering the metabolism of undecaprenyl-phosphate: the bacterial cell-wall unit carrier at the membrane frontier.Microbial Drug Resistance 20:199–214DOI 10.1089/mdr.2014.0035.

MartinW, Russell MJ. 2003. On the origins of cells: a hypothesis for the evolutionarytransitions from abiotic geochemistry to chemoautotrophic prokaryotes, and fromprokaryotes to nucleated cells. Philosophical Transactions of the Royal Society ofLondon B: Biological Sciences 358:59–83 DOI 10.1098/rstb.2002.1183.

Meyer BH, Albers S-V. 2013.Hot and sweet: protein glycosylation in Crenarchaeota.Biochemical Society Transactions 41:384–392 DOI 10.1042/BST20120296.

Miller MA, PfeifferW, Schwartz T. 2010. Creating the CIPRES science gateway forinference of large phylogenetic trees. In: Proceedings of the Gateway ComputingEnvironments Workshop (GCE). New Orleans, 1–8.

Mohammadi T, Karczmarek A, Crouvoisier M, Bouhss A, Mengin-Lecreulx D, DenBlaauwen T. 2007. The essential peptidoglycan glycosyltransferase MurG forms acomplex with proteins involved in lateral envelope growth as well as with proteinsinvolved in cell division in Escherichia coli.Molecular Microbiology 65:1106–1121DOI 10.1111/j.1365-2958.2007.05851.x.

Naparstek S, Guan Z, Eichler J. 2012. A predicted geranylgeranyl reductase reduces theω-position isoprene of dolichol phosphate in the halophilic archaeon, Haloferax vol-canii. Biochimica et Biophysica Acta 1821:923–933 DOI 10.1016/j.bbalip.2012.03.002.

Ogura K, Koyama T. 1998. Enzymatic aspects of isoprenoid chain elongation. ChemicalReviews 98:1263–1276 DOI 10.1021/cr9600464.

Penn O, Privman E, Ashkenazy H, Landan G, Graur D, Pupko T. 2010. GUIDANC: a web server for assessing alignment confidence scores. Nucleic Acids Research38:W23–W28 DOI 10.1093/nar/gkq443.

Peretó J, López-García P, Moreira D. 2004. Ancestral lipid biosynthesis and early mem-brane evolution. Trends in Biochemical Sciences 29:469–477DOI 10.1016/j.tibs.2004.07.002.

Lombard (2016), PeerJ, DOI 10.7717/peerj.2626 13/15

Petitjean C, Deschamps P, López-García P, Moreira D. 2015. Rooting the domainarchaea by phylogenomic analysis supports the foundation of the new kingdomProteoarchaeota. Genome Biology and Evolution 7:191–204 DOI 10.1093/gbe/evu274.

Philippe H. 1993.MUST, a computer package of management utilities for sequences andtrees. Nucleic Acids Research 21:5264–5272 DOI 10.1093/nar/21.22.5264.

Price MN, Dehal PS, Arkin AP. 2010. FastTree 2–approximately maximum-likelihoodtrees for large alignments. PLoS ONE 5:e9490 DOI 10.1371/journal.pone.0009490.

Ronquist F, TeslenkoM, Van der Mark P, Ayres DL, Darling A, Höhna S, LargetB, Liu L. 2012.MrBayes 3.2: efficient Bayesian phylogenetic inference andmodel choice across a large model space. Systematic Biology 61:539–542DOI 10.1093/sysbio/sys029.

Sanyal S, Menon AK. 2010. Stereoselective transbilayer translocation of mannosylphosphoryl dolichol by an endoplasmic reticulum flippase. Proceedings of theNational Academy of Sciences of the United States of America 107:11289–11294DOI 10.1073/pnas.1002408107.

Sato S, MurakamiM, Yoshimura T, Hemmi H. 2008. Specific partial reduction ofgeranylgeranyl diphosphate by an enzyme from the thermoacidophilic archaeonSulfolobus acidocaldarius yields a reactive prenyl donor, not a dead-end product.Journal of Bacteriology 190:3923–3929 DOI 10.1128/JB.00082-08.

Schenk B, Fernandez F, Waechter CJ. 2001. The ins(ide) and outs(ide) of dolichylphosphate biosynthesis and recycling in the endoplasmic reticulum. Glycobiology11:61R–79R DOI 10.1093/glycob/11.5.61R.

Shimizu N, Koyama T, Ogura K. 1998.Molecular cloning, expression and purifi-cation of undecaprenyl diphosphate synthase. Journal of Biological Chemistry273:19476–19481 DOI 10.1074/jbc.273.31.19476.

Sojo V, Pomiankowski A, Lane N. 2014. A bioenergetic basis for membrane divergencein archaea and bacteria. PLoS Biology 12:e1001926DOI 10.1371/journal.pbio.1001926.

Sparrow CP, Raetz CR. 1985. Purification and properties of the membrane-boundCDP-diglyceride synthetase from Escherichia coli. Journal of Biological Chemistry260:12084–12091.

Stamatakis A. 2014. RAxMLversion 8: a tool for phylogenetic analysis and post-analysisof large phylogenies. Bioinformatics 30:1312–1313DOI 10.1093/bioinformatics/btu033.

Stukey J, Carman GM. 1997. Identification of a novel phosphatase sequence motif.Protein Science 6:469–472.

Teo A.CK, Roper DI. 2015. Core teps of membrane-bound peptidoglycan biosyn-thesis: Recent advances, insight and opportunities. Antibiotics 4:495–520DOI 10.3390/antibiotics4040495.

Touzé T, Blanot D, Mengin-Lecreulx D. 2008. Substrate specificity and membranetopology of Escherichia coli PgpB, an undecaprenyl pyrophosphate phosphatase.Journal of Biological Chemistry 283:16573–16583 DOI 10.1074/jbc.M800394200.

Lombard (2016), PeerJ, DOI 10.7717/peerj.2626 14/15

Vranová E, Coman D, GruissemW. 2013. Network analysis of the MVA and MEPpathways for isoprenoid synthesis. Annual Review of Plant Biology 64:665–700DOI 10.1146/annurev-arplant-050312-120116.

Wächtershäuser G. 2003. From pre-cells to Eukarya—a tale of two lipids.MolecularMicrobiology 47:13–22.

Lombard (2016), PeerJ, DOI 10.7717/peerj.2626 15/15